Digital Karma Web Standards & Architecture

Digital Karma Federation Protocol v8.2

The Digital Karma Federation is an open standard for making websites easier for AI systems to discover, understand, and trust. Sites publish standardized endpoints, keep their discovery files in predictable root locations, and expose transparent relationship and quality signals. DigitalKarmaWeb.com is the specification authority, but any site can implement the standard independently.

Core Principles

1. Discoverability - Sites expose predictable machine-readable entry points.
2. Interoperability - JSON payloads and Schema.org markup use consistent patterns.
3. Transparency - Quality scores and compliance signals are openly documented.
4. Decentralization - Peer-to-peer discovery works without mandatory central approval.
5. Root Integrity - robots.txt and sitemap.xml stay at the site root.
6. Version Discipline - The published version must match the actual implementation.

What Changed in v8.2

Endpoint files are no longer enough to prove compliance. Every canonical sitemap page must render the site's DataCatalog and complete catalog Dataset graph, and a generated full-sitemap receipt must prove the result.

What Changed in v8.1 (previous revision)

v8.1 is a small, backward-compatible revision. AI crawlers kept requesting /llm.json after v8.0 retired it, so the file came back, this time generated from the manifest on every build instead of kept by hand. A v8.0 site is still a valid member.

What Changed in v8.0 (previous revision)

v8.0 is an architectural revision, not a terminology change. See the full changelog for the release history.

What Changed in v7.1 (previous revision)

Why v8.2 Is Worth Implementing

• It reduces ambiguity. The site can describe itself directly instead of relying on AI systems to infer intent from design and prose alone.
• It improves reuse. Datasets, endpoints, and summaries become easier for assistants and search systems to cite and traverse.
• It supports portfolio growth. Related sites can share a repeatable structure without collapsing into one hard-coded platform.
• It creates a real audit trail. Owners can validate what exists, what is missing, and what drifted after later edits.

Compliance Levels

Read Full Federation Spec

Required Artifacts

A v8.2 site publishes 6 required JSON files under /ai/ plus /llm.txt at the root.

Root Discovery Rules

Root placement remains mandatory. These are required operational discovery files, but they are separate from the seven public federation artifacts.

/robots.txt

• Always store it at the site root.
• Do not relocate it into /ai/.
• Allow intended public AI discovery paths.
• Include a Sitemap: reference to the root /sitemap.xml.

/sitemap.xml

• Always store it at the site root.
• Treat it as the canonical crawl sitemap.
• If it is generated automatically, update the generator rather than hand-editing the artifact.

/ai/sitemap.json

• Optional companion inventory for machine-readable content maps.
• Helpful, but not a replacement for the root XML sitemap.

Manifest and Health Examples

Manifest Example

{
  "protocol_version": "8.2",
  "supported_since": "2026-09-23",
  "deprecated_after": null,
  "removed_after": null,
  "name": "Your Site Name",
  "url": "https://yoursite.com",
  "federation_version": "8.2",
  "updated_utc": "2026-09-23T00:00:00Z",
  "endpoints": {
    "manifest": "/ai/manifest.json",
    "health": "/ai/health.json",
    "catalog": "/ai/catalog.json",
    "karma": "/ai/karma.json",
    "federation": "/ai/federation.json",
    "rendered_schema": "/ai/rendered-schema.json",
    "llm_txt": "/llm.txt",
    "llm_json": "/llm.json",
    "llm_json_ai": "/ai/llm.json",
    "llms_txt": "/llms.txt",
    "sitemap": "/sitemap.xml"
  },
  "related_sites": ["https://www.aiwebsitesystems.com"]
}

Health Example

{
  "protocol_version": "8.2",
  "supported_since": "2026-09-23",
  "deprecated_after": null,
  "removed_after": null,
  "site": "https://yoursite.com",
  "status": "ok",
  "federation_version": "8.2",
  "updated_utc": "2026-09-23T00:00:00Z",
  "last_rebuild_utc": "2026-09-23T00:00:00Z",
  "metrics": {
    "pageCount": 20,
    "datasetCount": 3,
    "entityCount": 4
  }
}

Schema and Scoring

Every canonical sitemap page must include parseable Schema.org JSON-LD with the publisher, website, primary page entity, one stable DataCatalog, and every Dataset in the site's catalog. The primary page entity must connect to the relevant Dataset layer.

Scoring Signals

Badge Thresholds

• Karma Certified at 0.70 or higher
• Karma Pro at 0.85 or higher
• Karma Elite at 0.95 or higher

Naming rule: Do not use Karma Bronze on v7.x sites. The correct name is Karma Certified.

Directory Blueprint

Minimum v8.2 Shape

/
|-- ai/
|   |-- manifest.json
|   |-- health.json
|   |-- catalog.json
|   |-- karma.json
|   |-- federation.json
|   `-- rendered-schema.json
|-- llm.txt
|-- robots.txt
|-- sitemap.xml
`-- datasets/                # optional but recommended

Pro / Elite Shape

/
|-- ai/
|   |-- manifest.json
|   |-- health.json
|   |-- catalog.json
|   |-- karma.json
|   |-- federation.json
|   |-- rendered-schema.json
|   |-- diagnostics.json
|   |-- sitemap.json
|   `-- journal/
|-- entities/
|-- federation/
|-- datasets/
|-- llm.txt
|-- robots.txt
`-- sitemap.xml

Validation Checklist

Before marking a site as v8.2 complete, verify all of the following:

• /ai/manifest.json loads and parses.
• /ai/health.json loads and parses.
• /ai/catalog.json loads and parses.
• /ai/karma.json loads and parses.
• /ai/federation.json loads and parses.
• /ai/rendered-schema.json reports a passing current audit.
• /llm.txt loads at the root.
• /robots.txt loads at the root.
• /sitemap.xml loads at the root.
• federation_version is "8.2" everywhere it appears, with protocol_version/lifecycle fields on every generated endpoint.
• Every canonical sitemap page exposes the complete catalog Dataset graph with stable identifiers.
• The primary page entity links to the relevant Dataset layer.
• The receipt hashes match the current sitemap and catalog.
• Constellation peer links are present.
• robots.txt points to the root sitemap.xml.
• No one relocated root discovery files into /ai/.